Condensation of Ammonia by Direct Contact

[lh1975]

We'd like to condensate a stream of ammonia vapour in a surge tank of subcooled liquid ammonia.
What equipment (sparger,...) should we use to inject the vapor inside the liquid ?
How can we determine the maximum quantity of vapor that we can inject to be sure that everything will be condensed ?
What are the design parameters to considered ?

Thanks
lh

 

[Montemayor]

lh:

You don’t give specific data on either the Ammonia vapor or the liquid, so I’ll have to resort to generalities.

The sparger you employ should be as efficient as you can buy or make. Sintered stainless steel in the form of a pipe would be my first choice. I’ve used it before and it worked very well when sparging the gas at the bottom of the receiver or tank. However, let’s be specific about the following:

The superheated ammonia vapor – as well as the supercooled ammonia liquid – have to be immaculately clean and free of any impurities, such as non-condensables, etc. In other words, you must ensure that the system obeys and follows the thermodynamic properties of PURE AMMONIA. Otherwise, I would not recommend you employ this system of cooling.

To determine the maximum quantity of NH3 superheated vapor that you can condense within the supercooled NH3 receiver or tank, you must make an accurate heat and mass balance around the system and the maximum quantity of superheated NH3 that you can put in it is dictated by the existing Maximum Allowable Working Pressure (MAWP) of your receiver or tank. In other words, what you are proposing is an adiabatic mix of these two fluids, with a resultant saturated NH3 liquid product. This is a BATCH operation and you cannot heat up the resultant saturated liquid any higher than that saturated temperature that corresponds to the saturated pressure equal to the MAWP of the receiver or tank. To do so, is to activate the NH3 safety relief valve that hopefully is installed and properly functioning on that same receiver or tank.

I have also used a similar technique on Ammonia 1st stage intercoolers on 2-stage Ammonia refrigeration cycles. The exceptions are that the end product is not a saturated liquid - but a saturated gas - and the process is continuous.

In order to make the heat and mass balance, you must accurately identify the quantity of supercooled NH3 liquid in the receiver or tank (your "heat sink") and be prepared to instantly shutoff the superheated vapor being sparged by using a sensitive and accurate pressure switch that is set at approximately 90% of the MAWP on the receiver or tank.

I would not run this operation without experienced engineering supervision.

I hope this helps answer your general questions.


Art Montemayor
Spring, TX

 

[25362]

My entry refers to the definition of "subcooled" ammonia. If we speak of a batch in a tank containing only ammonia the pressure should correspond to the liquid temperature.

If so, there is no "subcooled" liquid, just saturated liquid ammonia, and any addition of superheated vapor would mean transferring heat to the liquid and mass to the vapor, and the pressure would, of course, rise.

Mr Montemayor, am I right ?

 

[25362]

To lh1975, in short, it appears that, under the mentioned conditions, no condensation by direct contact will take place, just vaporization of the liquid accompanied by a loss of sensible heat of the superheated vapor.

 

[Montemayor]

25362:

You are absolutely right - and very observant. There can be no subcooled liquid NH3 when the enclosed liquid in a tank is allowed to attain thermal equilibrium with the usual ambient surroundings - such as in a conventional NH3 receiver.

However, what I suspect lh1975 has up his sleeve (as usual, we get little or no detailed data or information about the application and I have to assume something) is that he proposes to "drop" or dump his residual gaseous NH3 "line pack" in his refrigeration cycle into the NH3 evaporator in order to store it there rather than venting it to the atmosphere. I did this all the time when running NH3 refrigeration cycles and confronted with a need to shut down and evacuate a segment of the process in order to do inspection, cleanup, or maintenance. All my major vessels were rated for the same MAWP as the main NH3 receiver in order to allow a shut down for a week or more. The MAWP on the evaporator, the intercooler, the condenser and the receiver allowed me to use these vessels as temporary storage in order to salvage the refrigerant.

Consequently, my TEMA CKU evaporator had essentially a subcooled liquid when I shut the unit down and introduced hot NH3 gas into it for subsequent condensation and storage within the evaporator. This way I avoided emmissions problems and NH3 waste. I had to contain and store the cold NH3 liquid anyway, so I simply used it as my internal process heat sink and conserved my gaseous NH3 line pack that way. This is what I believe is at the root of the query here. But as you can imagine, although a very simple thermodynamic adiabatic condensation takes place, it can go awry or out of control if not carefully operated and supervised.

I hope I've been able to explain the technique in a clear and understandable manner.


Art Montemayor
Spring, TX

 

[25362]

To Mr Montemayor, from my own experience in ammonia refrigeration systems, if we assume the liquid is saturated at the prevailing pressure, any bubbling of superheated vapor would mean a drop in its sensible heat to provide latent heat.

Considering the large difference between latent and sensible heat -especially at evaporator pressures- I assume some (not much) liquid would evaporate, and no "equilibrium" vapor condensation would be expected. Am I right ?

 

[Montemayor]

25362:

My experience has shown me that you are correct in your statements. However, what I always found in applying the technique is that the initial surge of hot, pressured NH3 gas (supplied by the NH3 compressors) would initially pressurize the evaporator (in my case) well beyond the normal 1.5 psig operating pressure. The initial gases would quickly blanket the vapor space over the initially cold, saturated liquid -- and in doing so, created a "subcooled" liquid. Once in the vapor space, the hotter gases take longer to cool down and eventually condense. Once all the introduced gas is condensed you are, as you correctly state, right back into the conventional saturated liquid state.



Art Montemayor
Spring, TX

 

[25362]

Art Montemayor: thanks for clarifying the subject. It is up to lh1975 to tell us whether this exercise was indeed their original intention.

As an interesting corollary: when reviewing direct liquid-vapor contacting we see there is simultaneous heat and mass transfer.

In your practical example both heat and mass are transferred in the same direction, from vapor to liquid, in which the heat transfer coefficient (HTC) is generally larger than would be expected from heat transfer alone.

If, on the other hand, the vapor is cooled down and the liquid evaporates, the HTC would be lower than the expected values for heat transfer alone.

For hot ammonia gas cooling (sensible heat, Cp~0.6 Btu/lb) and complete condensation, the major heat load is the condensation, more than 500 Btu/lb, that must be supplied by the sensible heating of the liquid, about 1 Btu/lb.

It appears that whether the hot gases would partially or totally condense or just cool down upon reaching a new equilibrium, depends on the relative amount of cold liquid to warm gases in the evaporator.

 

[25362]

Cp units and sensible heat are, of course, per deg F.

 

[unclesyd]

I think lh1975 definitely needs to read the post and provide further information.
Not being that well versed in NH3 Storage I hope he isn't talking about cryogenic NH3 storage.

If I recall correctly something similar was the cause of a rollover and lost of containment in a cryogenic storage tank. This cause us to jump through hoops on our installation.

 

[lh1975]

I need to give you more information. Montemayor is very close to the exact case.
What we assume is a continuous cryogenic process (ammonia liquid near -28ºF. The surge tank where we want to condense the residual vapor operates at a few bars, so ammonia liquid is subcooled. There is no equilibrium in the tank. The pressure is maintained by injection of a portion of the residual ammonia vapor in the vapor phase. The rest of the residual ammonia vapor must be bubbled through the liquid to condense the vapor.\
I'd still like to know the maximum quantity of gas condensed into the liquid and how to design the injection points ?
(This case is close to what you could find with LNG).

 

[25362]

Lh1975: being your process a continuous one, it means chilled liquid ammonia is circulating at a prescribed rate. Right ?

You must decide what approach between the leaving -still subcooled- liquid ammonia temperature and that of the condensing vapors at the prevailing pressure, you wish to have. Data is scarce, but for water in barometric condensers it is about 3^oC.

The temperature difference "driving force" should be sufficiently large; for steam direct condensation with water, free of non-condensables, it lies between 10 and 20^oC.

I've made some quick calcs., using thermal data from ammonia tables, based on criteria similar to those for a steam/water barometric condenser free of non-condensables assuming a vessel pressure of 61 psia (BP: 31^oF), with slightly superheated vapors, and got a mass ratio of liquid ammonia to condensing vapors of about 10:1.

A wider temperature approach would mean a larger ratio.

Would this result be in the ballpark ? I wonder...

I suggest you read again Art Montemayor's first post in this thread.